Method, Device, and Program Product for Managing Computing System

2. The method according to claim 1, wherein the allocation model is generated based on existing knowledge used to allocate a set of operations to multiple computing units.

3. The method according to claim 1, wherein the predetermined condition includes: the direction of the reward is opposite to the direction of change in the performance index.

5. The method according to claim 1, wherein receiving the adjustment includes: receiving the adjustment from a user managing the computing system, and wherein the adjustment further includes an adjustment for an action space of the allocation model.

6. The method according to claim 1, further including: generating, based on the reward and in response to determining that the match degree does not satisfy the predetermined condition, training data for updating the allocation model.

8. The method according to claim 7, further including: updating the allocation model using the training data.

10. The method according to claim 1, wherein the computing units include graphics processing units in the computing system.

12. The electronic device according to claim 11, wherein the allocation model is generated based on existing knowledge used to allocate a set of operations to multiple computing units.

13. The electronic device according to claim 11, wherein the predetermined condition includes: the direction of the reward is opposite to the direction of change in the performance index.

15. The electronic device according to claim 11, wherein receiving the adjustment includes: receiving the adjustment from a user managing the computing system, and wherein the adjustment further includes an adjustment for an action space of the allocation model.

16. The electronic device according to claim 11, wherein the method further includes: generating, based on the reward and in response to determining that the match degree does not satisfy the predetermined condition, training data for updating the allocation model.